统计样本
Stat Sample
A stat sample is a collection of hands or observations used to calculate a player's statistical tendencies in poker tracking software.
Overview
In online poker, stat samples are the foundation of player profiling. A stat sample consists of a specific number of hands (e.g., 100, 500, 1000) that have been observed for a given opponent. Tracking software like PokerTracker or Hold'em Manager aggregates these hands to produce statistics such as VPIP (Voluntarily Put Money In Pot), PFR (Preflop Raise), and 3-bet percentage.
Importance of Sample Size
The reliability of a stat increases with sample size. A small sample (e.g., 20 hands) can be misleading due to variance. For example, a player who has raised 10 out of 20 hands (50% PFR) is likely not a 50% raiser long-term—they simply ran hot over a few hands. A sample of 500+ hands is generally considered more reliable, though even then, certain stats (like 3-bet percentage) require larger samples to converge.
Usage in HUDs
Most HUDs display stats with a sample size indicator (e.g., "56 hands"). Players should adjust their trust in a stat based on the sample. For common stats, a rule of thumb is:
- VPIP/PFR: converge after ~200-300 hands
- 3-bet: converge after ~500-1000 hands
- Fold to 3-bet: ~500 hands
Contextual Considerations
Stat samples are not just about quantity but also context. Stats from full-ring games may not transfer to 6-max. Similarly, a sample from a specific stakes or game type (e.g., tournaments vs. cash) should be treated separately. Some software allows filtering by position, stack size, or blind level.
Limitations
Regulation and player pool changes can affect stat reliability. A stat sample from 6 months ago may be obsolete if the player has improved. Additionally, sample size thresholds vary by stat; rarer events (e.g., 4-bet bluff) require larger samples. Always consider the sample alongside other information like hand history review and live reads.
Best Practices
- Use filters to isolate relevant samples (e.g., only 6-max cash hands).
- When sample is below 100 hands, rely more on general population tendencies.
- Update your database regularly to keep samples fresh.
- In fast-fold games (e.g., Zoom), prioritize sample size over recency due to rapid hand accumulation.